prisl-code 1.0.0

Autonomous CLI powered by local LLMs with self-healing tools

Prisl Code

A local AI coding assistant for your terminal that runs GGUF models using llama.cpp locally. It can read, write, and execute code in your projects while running offline in an isolated environment.

You can run your own server too it will detect it and will connect.

Installation

Prisl Code is designed to be a one-command setup.

pip install prisl-code

Usage

Once installed, navigate to any folder on your computer and start the assistant by typing:

prisl-code

On its first run, Prisl Code checks for an active LLM server (ports 8080/11434). If none is found, it offers to automatically download llama-server and opens a native file picker for you to select your GGUF model.

Basics

Add context:

• Type @filename.py anywhere in your message to instantly inject that file's contents into the AI's memory.

Commands:

• @<filepath>: Inject a file's content directly into context.
• /compact: Clear conversation history to save tokens (keeps system prompt).
• /history: Print a summary of the current conversation history.
• /save: Export the current chat history to a Markdown file.
• /clear: Clear the terminal screen.
• /help: Show this help menu.
• /exit: Exit the application.

Requirements

• Python 3.8+
• Pip 23.0+

Uninstall

pip uninstall prisl-code

License

This project is open source and available under the MIT License.

_{Developed by rx76d}